PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of clinbiorevLink to Publisher's site
 
Clin Biochem Rev. 2003 November; 24(4): 131–133.
PMCID: PMC1853352

Comments on the Proposed National Guidelines for Analysis of Cerebrospinal Fluid for Bilirubin in Suspected Subarachnoid Haemorrhage. Produced by a working group of UK NEQAS for Immunochemistry and in conjunction with the National Audit Group of the Association of Clinical Biochemists

Reviewed by Ainsley H Chalmers

Ann Clin Biochem 2003;40:481–8 

This is a useful document in that it outlines a detailed laboratory approach for the spectrophotometric detection of pigments in cerebrospinal fluid (CSF) which may indicate the presence of a subarachnoid haemorrhage (SAH). It should be emphasized that it is important to detect an SAH early as it may be either fatal or result in mild to serious neurological outcomes not to do so.1 Additionally, early surgery to repair aneurysmal bleeding and vasospasm results in improved clinical outcomes for the patient.1

As the article has indicated, computer tomography (CT) scanning is useful in the earlier events of a bleed, detecting over 90% of SAHs if used within the first 24 hours falling to 50% after 1 week post-ictus. A positive CT scan is then followed by cerebral angiography to confirm the site and presence of the bleeding aneurysm. However, the CT scan can become negative if a patient presents one or more days after a bleed commences. In these circumstances the spectrophotometric identification of CSF pigments becomes useful since it detects 100% of patients with SAH from 12 hours to 2 weeks post-ictus, falling to >70% after 3 weeks and 40% by week 4.2

The pigments which give CSF its slight discolouration in patients is due mainly to oxyhaemoglobin (HbO2), methaemoglobin (metHb) and bilirubin. After aneurysmal haemorrhage, erythrocytes rapidly disseminate throughout the subarachnoid space where they are lysed, with HbO2 being released and enzymatically metabolised to bilirubin. HbO2 occurs a few hours after a bleed but bilirubin takes about 12 hours to appear because of the time taken in metabolizing HbO2 to bilirubin. The term “xanthochromia test” has been applied to the spectrophotometric investigation of CSF pigments and will be referred to as such. Xanthochromia is derived from Greek, with xanthos meaning light coloured or blonde and chromia, colour. The term “xanthochromia positive” implies the presence of these pigments at a level sufficiently high to indicate the possibility of an SAH. Conversely “xanthochromia negative” means that the level of these pigments is sufficiently low to be of any significant clinical consequence. Different laboratories have placed different emphases on these haemoglobin-derived compounds in assessing whether a patient is xanthochromia positive or negative and this will be discussed more fully in this article.

Figure 2 of the UK guidelines shows a useful flow chart detailing the overall laboratory approach to patients with suspected SAH which may help in clinical diagnosis. However some comments need to be made about this approach and also to consider other aspects pertinent to these guidelines.

Traumatic Lumbar Puncture or Tap

A traumatic tap confuses the interpretation of this test because one is uncertain whether the pigments are derived from the tap (i.e. are artefactual) or are from an SAH. To overcome this problem usually 3 or more CSF fractions are collected. The guidelines suggest using the fourth fraction for spectrophotometric analysis with the rationale being that if there is a traumatic tap then the blood pigments/cells will be diluted out in the latter fractions of the tap. In our laboratory we use the third fraction of the tap. Once when we did the test on the first fraction a marginally positive result was obtained, however the result was negative when the third fraction was analysed (only the negative result was reported). The problem of the traumatic tap has not yet been resolved fully.

In our studies, we have found that a net bilirubin absorbance (NBA) of 0.042 absorbance units (AU) at 476 nm is equivalent to 1 μmol bilirubin/L and this has been confirmed in the guidelines. We also have found that 1 g Hb/L gives an absorbance at 415 nm of about 8 AU. If one considers a hypothetical situation in which there is a traumatic tap where 1% of the CSF is contaminated with blood from the vascular system and 0.01% of the Hb is lysed with the patient having a serum bilirubin of 10 μmol/L, an haematocrit of 0.5 and Hb of 120 g/L, then the following calculation can be made. The NBA will be 0.0021 AU (negative) and the net HbO2 absorbance at 415 nm (NOA) will be 0.096 AU. This calculation illustrates how critical the effect of erythrocyte lysis and subsequent HbO2 release is on the interpretation in the event of a traumatic tap whereas bilirubin only becomes a serious contamination at 1% when the concentration is about 40 μmol/L.

Perhaps in the event of a positive xanthochromia result, one solution would be to scan fractions 2 to 4 to see if they yield either similar or dissimilar spectrophotometric scans. Similar scans would suggest a non-traumatic tap and useful results and vice versa. If one uses exclusively only the third or fourth fractions there is always a lingering doubt that the positive result may be due to trauma and not an SAH.

The Diagnostic Usefulness of Net Absorbance Due to HbO2 versus Bilirubin

The article considers the possible roles of the three main pigments namely HbO2 (415 nm), metHb (403–410 nm) and bilirubin (460 nm) in the xanthochromia test. In our experience of over 5 years with this test, we have yet to see a CSF exclusively with metHb present, possibly because any HbO2 present hides its presence. Also, if metHb is present it is not known what absorbance change at 403–410 nm indicates a xanthochromia positive result. Some consider NOA due to HbO2of 0.023 AU or greater with or without the presence of bilirubin as being diagnostic of an SAH.3 In our initial study we found that the NOAs were no different in the patients with confirmed SAH to those without SAH.4 The guidelines suggest using exclusively the NBA at 476 nm because of the possibility that the trauma associated with the tap would generate HbO2. The guidelines also showed that HbO2 is a poor predictor of SAH. Nonetheless they suggest that specimens should not be processed through a pneumatic tube or be spun gently at about 1,000 g for 5 min (we microfuge for 1 min). We find the comment that HbO2 can interfere with the NBA is incorrect because at the wavelength of measurement, namely 476 nm, NOAs up to 2 AU give a zero NBA and hence do not interfere with the xanthochromia test.5 However, perhaps different spectrophotometers show an NOA interference with the 476 nm reading. Obviously then the wavelength at 476 nm may need to be changed slightly to overcome this interference. Also in CSF, to which has been added erythrocytes in vitro, there is no observable conversion of HbO2 to bilirubin at room temperature over 48 hours (A.H. Chalmers, unpublished observation).

Reference Range

The guidelines use a cut off of an NBA of 0.007 AU at 476 nm wavelength, which we have recently adopted. Where the patient has a positive xanthochromia result (i.e. NBA >0.007) and serum bilirubin >20 μmol/L, the guidelines recommend a calculation to determine the contribution of serum bilirubin to the NBA of CSF as follows:

equation image

We avoid this PA calculation because it is an approximation which tends to overestimate the serum bilirubin influence on CSF bilirubin by up to 15%. It has been recommended that albumin instead of total protein concentration ratios be used for greater accuracy but even still this results in an overestimation.6,7 It is currently not known what the relationship between elevated serum bilirubin and CSF bilirubin is in patients without SAH. This experiment would be difficult to do because of the significant influence of even minor traumatic taps on the NBA result in patients with hyperbilirubinemia. However, if done properly it would allow PA to be determined more accurately. In our first paper on the test we found that the correlation in 31 patients between CSF bilirubin and serum bilirubin was very poor with r = −0.058 but in this study serum bilirubin was close to the normal range and ranged from 5–30 μmol/L.4 A serious consequence of this calculation is that it has the potential to generate a false negative result and thereby give rise to a worse clinical scenario in sending home a patient who has a bleeding aneurysm. A false positive result would result in a lesser evil being done, that is, a cerebral angiogram which has a milder associated risk.

A reference range of 0.007 AU translates to a bilirubin concentration of 0.17 μmol/L. There are indications in Australia that certain laboratories are attempting to estimate CSF bilirubin using automated high throughput analyzer systems. My understanding is that sensitivity with these is close to 0.6 μmol bilirubin/L which is far above the reference range. For these instruments to be useful they would need a sensitivity about 0.1 μmol/L and be able to measure accurately at the cut off level of 0.17 μmol/L. It is obvious from this that an external QC program is needed so that laboratories can assess their methodologies. From discussions with others I believe that this is currently happening in Australia. In the UK there is already such a QC program in place (R. Beetham, personal communication).

Other Aspects

The other area in which we differ to the guidelines is in the dilution of CSF. The CSF volumes we receive are about 100–300 μL and not as ample as those recommended in the guidelines (a minimum of 1 mL). For this reason we scan in a 100 μL cuvette. Occasionally when the specimen volume is only 50–80 μL, before scanning, we dilute the specimen 1:1 with water (but no more) and double the final result.

We also do not factor in the CSF protein in our decision making because the clinicians should be aware that their patient may have other clinical conditions influencing CSF protein concentration, such as cerebral infections, and secondly because the time constraints in doing this estimation and the semi-urgent nature of the xanthochromia test.

The guidelines are suspicious of HbO2 absent with only bilirubin detected up to 1 week post-ictus. Caution should be exercised here since the rate of HbO2 conversion to bilirubin intracranially is currently not known. Also the presence or absence of HbO2 may be dependent not only on the time post-ictus but also on the extent of the bleed, that is, whether it is a small or large bleed.

The scans shown in the guidelines stop at about 370 nm. We recommend scanning to 350 nm to obtain a more accurate tangent skim. The guidelines also recommend a full-scale deflection (FSD) of >0.1 AU. We autoscale our scans, sometimes to 0.02 FSD4, so that they fill the whole page of our printout thereby allowing a more accurate assessment of NBA.

In conclusion, the test as outlined in the guidelines is useful in confirming SAH where the CT and possibly the cerebral angiogram are negative. A recent study has shown that the test as outlined in the guidelines performs well with a sensitivity and specificity of 0.8 and 1.0 respectively.8 The problem of contamination from a traumatic tap is ever present and one should be vigilant about this. The use of this test in patients with hyperbilirubinemia (including neonates) and other conditions associated with a leaky blood brain barrier should be treated with caution or better still be avoided.9 Also in our experience, the results cannot be used in diagnosing SAH in patients with cerebral infections/inflammatory processes such as meningitis or in patients with brain tumours.

Footnotes

The guidelines may also be located at http://www.immqas.org.uk/csfguidelines.pdf

References

1. Edlow JA, Caplan LR. Avoiding pitfalls in the diagnosis of subarachnoid hemorrhage. New Eng J Med. 2000;342:29–36. [PubMed]
2. Vermeulen M, Van Gijn J. The diagnosis of sub-arachnoid haemorrhage. J Neurol Neurosurg Psychiat. 1990;53:365–72. [PMC free article] [PubMed]
3. Cruickshank AM. CSF spectrometry in the diagnosis of subarachnoid haemorrhage. J Clin Path. 2001;54:827–30. [PMC free article] [PubMed]
4. Chalmers AH, Kiley M. Detection of xanthochromia in cerebrospinal fluid. Clin Chem (Letter) 1998;44:1740–2.
5. Chalmers AH. Cerebrospinal fluid xanthochromia testing simplified. Clin Chem (Letter) 2001;47:147–8.
6. Kjelin KG. The binding of xanthochromic compounds in the cerebrospinal fluid. J Neurol Sci. 1969;9:597–601. [PubMed]
7. Wahlgren NG, Bergstrom K. Determination of haem derivatives in the cerebrospinal fluid – a semi-quantitative method. J Neurol Neurosurg Psychiatry. 1983;46:653–8. [PMC free article] [PubMed]
8. O’Connell DM, Watson ID. Definitive angiographic detection of subarachnoid haemorrhage compared with laboratory assessment of intracranial bleed in CT-negative patients. Ann Clin Biochem. 2003;40:269–73. [PubMed]
9. Hansen TW. Mechanisms of bilirubin toxicity:clinical implications. Clin Perinatol. 2002;29:765–78. [PubMed]

Articles from The Clinical Biochemist Reviews are provided here courtesy of The Australian Association of Clinical Biochemists